Dynamics Object Tracking for Mobile Robots in Indoor Environments with Crossing Path

This paper proposes an obstacle-tracking technique that uses a modified data filter in the Probabilistic Data Asso-ciation Filter (PDAF) with an Ensemble Kalman Filter (EnKF). This technique improves tracking and prediction accuracy, es-pecially in crowded and dynamic indoor environments. Indoor environments were designed to include many items, such as pedestrians, human interaction with static and moving objects, walls, and other robots. Several objects that move linearly from an initial point to an endpoint are considered. However, some objects were created to cross paths and perform intersection maneuvers to evaluate the algorithm's ability to track objects and avoid object prediction errors. Three scenarios were used to test the PDAF algorithm, including the crossing route of an object. Several conditions present substantial challenges to traditional tracking algorithms. Multiple scenarios were tested to determine the adaptability and effectiveness of the PDAF algorithm. The algorithm's ability to reliably forecast the movements of these items was measured using the Root Mean Square Relative Error (RMSRE) metric.